You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the bug
A user in discord reported this. In their logs, ffmpeg took 3 mins to cut the video. They pulled the boss before this time elapsed and the app fell over.
Expected behavior
The ffmpeg cutting the video should be async in respect to the recorder process. Currently, the recording buffer does not restart until the cutting is done. That's bad logic. See the code in recorder.stop, particularly await this.finalizeVideo(metadata, outputFilename);.
Quoting from Discord
huh this is an interesting one and is definitely a bug in our code
the problem seems to be that it took your PC ~3 mins to cut one of your pantheon pulls to size
in that time, the app isn't ready to start another recording, but you pulled again in that time
so it fell over
as for why it took so long to cut the video to size, that's very strange and I'm not sure. Was your PC super busy? Anything notable going on?
Regardless of how long it takes, we shouldn't require it completes before the next pull so there is some code fixes I can make to help this.
The cutting of the video usually takes 2-3 seconds in normal circumstances so probably just never hit this before as no one manages to pull within that window.
Things all go downhill from here, as we previously cancelled the recording buffer (which was actually a no-op), so we get a buffer restart inside a genuine recording. Full log is attached.
The video cutting is consistently slow but only causes a problem when a following pull happens during the cut (so if they take 10 mins on trash its fine, but if they wipe and quick re-pull its not).
Describe the bug
A user in discord reported this. In their logs, ffmpeg took 3 mins to cut the video. They pulled the boss before this time elapsed and the app fell over.
Expected behavior
The ffmpeg cutting the video should be async in respect to the recorder process. Currently, the recording buffer does not restart until the cutting is done. That's bad logic. See the code in recorder.stop, particularly
await this.finalizeVideo(metadata, outputFilename);
.Quoting from Discord
Things all go downhill from here, as we previously cancelled the recording buffer (which was actually a no-op), so we get a buffer restart inside a genuine recording. Full log is attached.
WarcraftRecorder-2022-10-03 (3).log
The text was updated successfully, but these errors were encountered: